Reply to Jobling, "Ectopic Expression of the ydaS and ydaT Genes of the Cryptic Prophage Rac of Escherichia coli K-12 May Be Toxic but Do They Really Encode Toxins?: a Case for Using Genetic Context To Understand Function".

REPLY

My group thanks Dr. Jobling for his insightful letter (1) that argues for a regulatory role for YdaS/T, encoded in the rac prophage of Escherichia coli. This argument is brought up as a critique of our suggestion that YdaS/T may be a toxin. We are glad that our work has generated interest in the community and take this opportunity to present the background to our study, reported in two papers, and also summarize our findings to remove any ambiguity that our papers might have caused in the minds of our readers.

Dr. Jobling is correct in pointing out that our paper focuses “solely on the toxicity of these two genes at the expense of any insight into their toxicity.” We also appreciate Dr. Jobling’s perspective, which arises from his own recent work on toxins encoded in full-length rac prophages encoded in pathogenic E. coli (2). Indeed, the fact that these proteins share homology—albeit limited at the sequence level—with lambda Cro and cII would throw insight into potential mechanisms of action. For example, it was shown that overexpression of cII causes toxicity by perturbing replication (3), a work that Dr. Jobling quotes. Similar analysis of the mechanism of toxicity of YdaS/T is certainly of interest, but that was not quite the objective of our study and is unlikely to be pursued by our lab in future.

Early in 2011, we were interested in knowing whether horizontally acquired transcription factors in E. coli K-12 had any effect on transcription of core genes. By chance, we landed upon RacR, a prophage-encoded putative transcription factor with weak homology to lambda cI. We realized that we could not delete the racR gene, and neither could the Keio collection effort (4). Our objective therefore was to find out what RacR’s targets were and how its function as a transcription factor might contribute to its essentiality to the E. coli cell. Our objective here was not to elucidate the effect of RacR or its targets on the life cycle of the rac prophage in its intact or defective form. Over the next few years, we gathered evidence to support the view that RacR represses ydaS/T and that this activity might be responsible for its essentiality. We summarize key pieces of evidence gathered from our two papers below:

Deletions of ydaS and ydaT fully suppress the lethality of ΔracR. ydaS and ydaT are not expressed under normal growth conditions. Across a variety of conditions sampled in publicly available transcriptome sequencing (RNA-seq) data sets, ydaS/T is expressed to a level comparable to that of the well-established silent bgl operon (5).

Ectopic expression of ydaS and ydaT from arabinose-inducible pBAD vectors is toxic to the cell (5). We were indeed worried about expression from a plasmid and artifacts that might arise from this construct. At this point, we established a collaboration with Devashish Rath, who had tools based on clustered regularly interspaced short palindromic repeat(s) (CRISPR) to downregulate racR and study its effect on ydaS/T expression.

CRISPR-mediated downregulation of racR increased the expression of ydaS/T (6).

Downregulation of racR also resulted in decreased fitness. However, downregulation of racR did not affect fitness in the ΔydaS-ydaT background (6).

RacR binds to the intergenic region between racR and ydaS, as measured by chromatin immunoprecipitation-quantitative PCR (ChIP-qPCR) and biochemical thermal shift assays and electrophoretic mobility shift assays (EMSAs) (5).

The rac prophage is mosaic across E. coli strains as shown by bioinformatic analysis of publicly available E. coli genomes. Even the occurrence of racR and ydaS/T is patchy. However, ydaS/T is present only in strains in which racR is also present (5).

Together, these show that the repression of ydaS/T by RacR is largely responsible for its essentiality to its host E. coli K-12 cell. These indicate that YdaS/T expression—from a plasmid, or from the chromosome when racR is downregulated—is toxic to the cell. The mechanism of action of YdaS and YdaT is not known, and their similarities to Cro and cII might give some pointers in that direction.

These results do not in any way argue against the potential role of these proteins as regulatory switches within the phage; addressing this issue was always beyond the scope of our work. It is indeed possible and quite likely that the toxic activity of these proteins against E. coli is incidental and a by-product of their regulatory activity within the rac prophage.